Nature douses dilution experiment.
An investigative team organized by NATURE reports in the July 28 issue that the results of a controversial dilution experiment published in the journal four weeks ago are "not reproducible in the ordinary meaning of that word." A repeat performance got four positive and three negative results at the University of Paris-Sud laboratory of Jacques Benveniste, who led the original 13 researchers.
In the original experiment, a solution of antibodies appeared to evoke a reaction from certain white blood cells and change their ability to hold a stain even after the antibody solution had undergone 120 tenfold dilutions (SN: 7/2/88, p.6). After a succession of so many dilutions, it is unlikely that even one molecule of antibody would be left in the solution. That scientists from six laboratories worldwide reported a reaction brought hosannas from many practitioners of homeopathic medicine, a 200-year-old practice that has endured criticism for its use of infinitesimal doses of drugs to stimulate a cure.
NATURE's skepticism was clinched by the final three runs of the experiment -- the only ones performed double-blind, meaning all test tubes had been randomly coded twice. The person measuring the cell's reaction to the antibodies could not have been influenced by a preconceived idea of the results. All three of these runs were negative.
One partially blind run of the experiment, however, produced positive results. For this, the mixtures of antibodies and cells were randomly transferred from the test tubes onto slides by one of the observers, Walter Stewart of the National Institutes of Health, well known for uncovering scietific fraud. "It was a blind reading," Stewrt explains, "but not a blind preparation of the experiment."
Also on hand were James Randi (known as The Amazing Randi, a former magician who has devoted the past 20 years to disproving pseudoscience), NATURE Editor John Maddox and Randi's assistant, Jose Alvarez.
In his editoral reply, Benveniste blasts NATURE for a "mockery of scientific inquiry." He maintains the last three tests "worked poorly mainly due to erratic controls." Also, Benveniste writes, the workload imposed by the week-long inquiry influenced the results of the final runs. The whole series of experiments was carried out primarily by Elizabeth Davenas, whose name appears first on the original paper.
"That's ridiculous," says Maddox, referring to the French scientist's explanation. "We were astonished at the atmosphere of the lab. They had done nothing to figure out why they had gotten these [original] results."
Also surprising to NATURE, he says, was that the first paper did not report that one of the coauthors from Israel had dropped out at the last minute in a dispute over the results, or acknowledge that a homeopathic drug company is a financial supporter of Benveniste's lab.
Benveniste says these facts distort the journal's report. "Does homeopathic companies paying two researchers ... mean that they order them into improper conduct?" he writes.
For his part, Randi says he could spot no trickery on the part of the researchers, although he did describe Davenas as a woman "very fond of rounding numbers," who recorded her data in pencil. "They just couldn't perform under pressure," he remarks. "That's the smoking gun."
The dilution research is not completely blown out of the water, however. NATURE will now begin closely examining data from other laboratories whose positive results were reported in the original paper, especially the one in Israel, which most strongly supports Benveniste's experiment. Says Maddox, "We shan't leave it alone."
|Printer friendly Cite/link Email Feedback|
|Date:||Jul 30, 1988|
|Previous Article:||Private parts = private property?|
|Next Article:||Chaos in a cold cloud of trapped ions.|